mozilla :: #datapipeline

21 Apr 2017
11:15gfritzschetrink: i'm not up to date on that one, chutten?
13:33trinkgfritzsche: there are 2 issues 1) there is a 20+% increase in release channel crash ping submissions 2) the average release crash ping size is now 84% bigger
13:36trinkgfritzsche: forwarded you the email thread
13:46trinkthe new latency mismatch match between the crash and main pings could account for the volume issue
13:55frankharter: re: pattern matching in map functions, think we should use case for single arg functions?
13:55frankI can't decide
13:56gfritzschetrink: looking over this in a bit
13:57harterI haven't given it too much thought, but it looks like the primary benefit of the case statement is disambiguating between (a, b) => c and ((a, b)) => c
13:57harterThere's no conflict for a => b though
13:57frankright, that's what I was thinking
14:21sunahsuhi have zero context, but: decomposition is much nicer to read than tuple accessors
15:05trinkgfritzsche: as you can see the monitors you requested are running in prod :)
15:08gfritzschetrink: which is great! :)
15:08gfritzscheknowing & understanding those events is what we need
15:11trinkyeah the only alert with diagnostics is ingestion_error
15:59trinkfrank: +1 on the duplicate analysis
17:01sunahsuhhmm, so, i tried to rerun this query and redash says it's running but it's not showing up in the presto dashboard at all, as far as I can tell:
17:01sunahsuhany idea what's going on?
17:01sunahsuhit's in query purgatory
17:29sunahsuhi haz theory -- taking it to #datatools
22:37ilanadoes anyone know off the top of their head the field from telemetry that populates search_counts in main_summary?
22:39frankilana: keyed histogram, search_counts
22:39ilanafrank, you're a star.
22:40frankaww, thanks :)
22 Apr 2017
No messages
Last message: 123 days and 6 hours ago